Anthropic mulls releasing new AI model

The Express Tribune reported, citing sources, that Anthropic is weighing a new AI model to counter the momentum of rival OpenAI since GPT-6 Astra arrived. The possible release comes as the company is expected to pursue a public listing, and follows Dario Amodei's plea that AI developers slow the speed of new capabilities over safety worries. Amodei set out that position in an essay on September 12.

Amodei's 3,800-word piece warned that swarms of AI agents could overrun the internet and slip beyond human control, and it drew backing from Sam Altman and Elon Musk. On September 3, OpenAI unveiled GPT-6 Astra, highlighting improvements in computer use, software engineering and cybersecurity, and the model is winning over businesses. Ramp, a corporate expense platform, put Astra at roughly 13% of tracked enterprise AI spending, while Anthropic's Claude Fable held 8%.

Revenue for Anthropic stood above $65 billion on an annualised run-rate basis as of July's close, compared with roughly $9 billion when 2025 ended. Its 2028 revenue projection sits between $190 billion and $200 billion, and OpenAI's comparable figure cleared $40 billion in July. Some investors still doubt that Astra poses a major threat to Anthropic, but open-source models that lower token costs are seen as a larger challenge.